investigation on pn formation at gdi engines at high loads · 2018. 7. 12. · institut fÜr...

www.kit.edu

INSTITUT FÜR KOLBENMASCHINEN Leiter Prof. Dr. sc. techn. Thomas Koch

KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft

Investigation on PN Formation at GDI Engines at High Loads

Injektor A

Denis Notheis (M.Sc.), Dr.–Ing Markus Bertsch, Dr.–Ing Amin Velji, Prof. Dr. sc. techn. Thomas Koch

2 20.06.2018 Institut für KolbenmaschinenLeiter Prof. Dr. sc. techn. Thomas Koch

Introduction

Source: Wiese et al.; Anforderungen an den Mehrlochinjektor zur Erfüllung zukünftigerEmissionsgrenzwerte beim direkteinspritzenden Ottomotor ; Int. Motorenkongress 2015

Previous research: focus of catalysator heating und moderate engine torque

At High loads the particle emissions are rising continuously! This research project

Now with Euro 6d-Temp 6 *1011 [#/km] in the

WLTP and RDE Small Particle size

challenge of GPF!

Source: Wurms et al.; Der neue Audi 2.0l Motor mit innovativem Rightsizing – ein weiterer Meilenstein der TFSI Technologie; Wiener Motorensymposium 2015

High loads became more important with new test cycles



Methodology

in cylinder optical investigation

Preinvestigations regarding spray and charge motion

CFD simulation of the mixture preparation

Exhaust gas primary particle number measurement



Content

Introduction

Experimental and numerical set-up

Results

Conclusion



p,T

Thermocouple

p

Charge amplifier

p,T

Indication System

Photo-multiplier

engine speed [rpm] 2000

IMEP [MPa] 1.4

Temperature oil, water [°C] 90

Temperature charge [°C] 30

relative air-fuel ratio [-] 1

Injector Lowflow Highflow

Type Solenoid actuated

Qstat [g/min] 620 820

Max. pressure [MPa] 50 35Basic Tumble Swirl

Measurement SetupEngine conditions and technical data

.



p,T

Thermocouple

p

Charge amplifier

p,T

Indication System

Photo-multiplier

camera La Vision HSS 6recording rate [kHz] 13.5 image chip [Pixel] 512 x 512 recording window [°CA aTDCf] -310 bis 90Light source Storz Technolight 270

displacement [cm³] 498 cm³stroke [mm] 90 mmbore [mm] 84 mmcompression ratio [-] 10.5:1setting range cam phasor [°CA] 80max. valve lift I/E [mm] 0.5-9.7 / 9.7

Measurement SetupOptical investigations in the engine

intakevalves

injector

spark plug exhaustvalves



v

v

Tumble 60 Basic

Calculation with StarCD© and es-ice© ~ 2 mio. cells

Simulation SetupMesh and Models

Flow fieldAir composition N2 ,O2 ,CO2 , H2O

Turbulence RANS k-ε /RNG Model

Spray

Multicomponent fueln-hexane (34 - % m)iso-octane (45 - % m)n-decane (21 - % m)

Primary breakup Distribution functionSecondary breakup Reitz and DiwakarWall - Impingement Bai – GosmanLiquid Film boiling Rosenow



Content

Introduction


Results

Conclusion



Parti

kela

nzah

l-Ko

nzen

tratio

n C

PC [#

/cm

³]

103

104

105

106

107

108

109

Einspritzbeginn [°KW vZOT ]220260300340

Basis-Injektor

Basis

103

104

105

106

107

108

109


Basis-Injektor

Tumble

103

104

105

106

107

108

109


Basis-Injektor

Drall

Basis-Injektor Reduzierter Durchfluss



Basis-Injektor Reduzierter Durchfluss Red. Durchfluss 500 bar



Parti

kela

nzah

l-Ko

nzen

tratio

n C

PC [#

/cm

³]

103

104

105

106

107

108

109


Basis-Injektor

Basis

103

104

105

106

107

108

109


Basis-Injektor

Tumble

103

104

105

106

107

108

109


Basis-Injektor

Drall




charge motion: reduction PN

Red. hydr. flow: Reduction of PN in

higher a higher variation of SOI

charge motion: significant reduction PN

higher load: increase PN

Red. hydr. flow:Increase PN

higherinjection pressure:

Reduction PN

WO

T (8

bar I

MEP

)Influencing factors on particle emissionstiming, charge motion and injection rate

Start of Injection [°CAbTDcf] Start of Injection [°CAbTDcf] Start of Injection [°CAbTDcf]

Start of Injection [°CAbTDcf] Start of Injection [°CAbTDcf]

Baseline SwirlTumble

part

icle

num

ber

conc

entr

atio

n C

PC [#

/cm

³]

char

ged

(14

bar I

MEP

)

Baseline

part

icle

num

ber

conc

entr

atio

n C

PC [#

/cm

³]

Start of Injection [°CAbTDcf]

Start of Injection [°CAbTDcf]Start of Injection [°CAbTDcf]

SwirlTumble

Start of Injection [°CAbTDcf] Start of Injection [°CAbTDcf]

Highflow 200 bar Lowflow 200 barLowflow 500 bar








v

320 280 240 200104

105

106

107

108 Baseline - Highflow - 350 bar Baseline - Lowflow - 350 bar

part

icle

num

ber

conc

entr

atio

n [#

/cm

³]

SOI [°CA bTDFf]

ResultsInfluence of injection rate on poolfire

Liquid film spot 1

0 2.5 5 7.5 10

Liquid film Thickness[µm]

Highflow 350 bar Lowflow 350 bar

More liquid film spots in the simulation of the highflow injector

High increase of particle number emission at early injection timings with the highflow injector



v

simulation at TDC optical investigation

Liquid film spot 1

ResultsInfluence of injection rate on poolfire

320 280 240 200104

105

106

107

108 Baseline - Highflow - 350 bar Baseline - Lowflow - 350 bar

part

icle

num

ber

conc

entr

atio

n [#

/cm

³]

SOI [°CA bTDFf]

High increase of particle number emission at early injection timings with the highflow injector

More liquid film spots in the simulation of the highflow injector

Optical investigation shows pool fire on the same spot as predicted by CFD simulation

Highflow 350 bar

I E

I E

Spark plug

Injector

View

Inlet

Inlet

Exhaust Exhaust



v

320 280 240 200104

105

106

107

108 Swirl - Lowflow - 200 bar

part

icle

num

ber

conc

entr

atio

n [#

/cm

³]

SOI [°CA bTDFf]

ResultsInfluence of large scale motion

0 1 2 3 4 5

Liquid film Thickness[µm]

SOI 280 SOI 240

560 °CA

high decrease of particle number emission from SOI of 280 to 240 °CA bTDCf

High impingement on one spot on the wall caused by the influence of swirl on spray targeting at late injection timings



ResultsInfluence of large scale motion

v

320 280 240 200104

105

106

107

108 Swirl - Lowflow - 200 bar

part

icle

num

ber

conc

entr

atio

n [#

/cm

³]

SOI [°CA bTDFf]

high decrease of particle number emission from SOI of 260 to 240 °CA bTDCf

High impingement on one spot on the wall caused by the influence of swirl on spray targeting at late injection timings

Optical investigation shows soot luminescence outwards from top land on the same spot as predicted by the CFD

Optical investigation

SOI 240

560 °CA 640 °CA 720 °CA



Results -Summary of highest influence factors



Results -Summary of highest influence factors

Large scale motion strongly affected the spray targeting.

With a Tumble Inlay the best mixture preparation was possible.

If the valve is been impinged the mixture preparation is influenced negatively.

High injection pressure influenced the evaporation rate of the spray and improve the mixture preparation if not influenced by the valves.

Colder wall temperature affect only the liquid mass.

With early injection timings piston impingement is increased otherwise with late injection timings mainly the liner is impinged.



Content

Introduction


Results

Conclusion



Conclusion

The potential of the different variation parameters to reduce the particle number emission is not additive. The variation parameters (valve overlap, tumble, rail pressure, …) affect each other mutually!

The simulation shows the trends of the mean particle measurement results, cycle-to-cycle variations could not be predicted with the used models

CFD-simulation predicted that the highest influencing factors on particle is the remaining liquid phase in the cylinder.



Thanks to…

Forschungsvereinigung Verbrennungskraftmaschinen e.V. (FVV, Frankfurt)Bundesministerium für Wirtschaft und Energie (BMWi) Arbeitsgemeinschaft industrieller Forschungsvereinigungen e.V. (AiF)Chairman Dr.-Ing Daniel Sabathil (Opel Automobile GmbH , Rüsselsheim)FVV working group, especially Bosch and DelphiSiemens PLM ICE-Support Team, especially Dr. Oleksiy Kochevskyy

and for your kind attention



Appendix: ValidationSpray and flow validation

0 1 2 3

0

25

50

75

100

time [ms]

0 1 2 3

0

50

100

150

200

time [ms]

pene

tratio

nve

loci

ty[m

/s]

pene

tratio

nde

pth

[mm

]Spray Validation in pressure chamber

Velocity[m/s]

Simulation

Measurement

Flow Validation in flowbench

SimulationMeasurement

+


investigation on pn formation at gdi engines at high loads · 2018. 7. 12. · institut fÜr...

Documents